Journal of Computer Languages最新文献

{"title":"Improving LLM-based code completion using LR parsing","authors":"Md Monir Ahammod Bin Atique ,&nbsp;Hyeon-Ah Moon ,&nbsp;Isao Sasano ,&nbsp;Kwanghoon Choi","doi":"10.1016/j.cola.2025.101352","DOIUrl":"10.1016/j.cola.2025.101352","url":null,"abstract":"<div><div>Code completion is a crucial feature in modern IDEs, improving programming efficiency. Traditional systems rely on prefix filtering and static ranking but often overwhelm users with lengthy, alphabetically sorted lists. Recent research has introduced LR-parsing-based approaches that derive completion candidates from language syntax and compute their ranks using open-source programs; however, these methods only suggest structural candidates, requiring manual refinement into complete code. To address this, we propose a hybrid method that integrates LR parsing with LLMs to enhance accuracy and usability. Our approach refines structural candidates using LR parsing into textual code suggestions via an LLM, referencing a database of ranked candidates from open-source programs. This combines the syntactic precision of LR parsing with the generative capabilities of LLMs. This study examines whether LLMs benefit from LR structural candidates in code completion. By comparing completions with and without these candidates, we assess their impact. Building on prior research, we also explore how leveraging top-ranked structural candidates can effectively enhance LLM-based code completion precision. We also demonstrate our method through VSCode extensions for Microsoft Small Basic and C. As a language-agnostic solution, our system applies to any language with a defined LR grammar. Our findings suggest that integrating LR parsing with LLM-based completion improves both accuracy and usability, paving the way for more effective code completion in modern IDEs.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"84 ","pages":"Article 101352"},"PeriodicalIF":1.8,"publicationDate":"2025-07-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144750140","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Novice modelers’ subjective comprehension and interaction with token-animated process models","authors":"Ilia Maslov ,&nbsp;Stephan Poelmans ,&nbsp;Yves Wautelet ,&nbsp;Frederik Gailly","doi":"10.1016/j.cola.2025.101350","DOIUrl":"10.1016/j.cola.2025.101350","url":null,"abstract":"<div><div>Process modeling is fundamental for effective (business) process management. Comprehension of process models by novice modelers and the effective integration of learning technologies present crucial challenges that can be addressed through the application of visualization, animation, and simulation techniques. In this study, we examine the experiences and perceptions of novice modelers deploying token-animated process models, drawing upon data from 119 college students specializing in business management and business engineering who answered comprehension questions based on these models. We concentrate on investigating perceived understanding through the utilization of Technology Adoption Model (TAM) constructs, employing Partial Least Square to validate an extended research model based on TAM. We additionally analyze qualitative data from respondents' answers to open questions to extract codes and themes that complete the research model's findings. The results confirm that token-animated process models are useful and preferred as a learning technique. Tokens enhance cognitive facilitation by incorporating visualization, animation, and simulation functionalities, resulting in improved objective and perceived comprehension. We extend the comprehension determinants of process models with perceived enjoyment and show that emotional states are also important in the utilization of tokens for teaching purposes. Over 80 % of participants reported a clear preference for using token-animated process models, confirming high levels of student acceptance. Our study also identified recommendations for enhancement and potential limitations associated with the use of animated tokens in education. Further theoretical and practical implications are finally discussed.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"84 ","pages":"Article 101350"},"PeriodicalIF":1.8,"publicationDate":"2025-07-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144757584","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"TestLoter: A logic-driven framework for automated unit test generation and error repair using large language models","authors":"Ruofan Yang,&nbsp;Xianghua Xu,&nbsp;Ran Wang","doi":"10.1016/j.cola.2025.101348","DOIUrl":"10.1016/j.cola.2025.101348","url":null,"abstract":"<div><div>Automated unit test generation is a critical technique for improving software quality and development efficiency. However, traditional methods often produce test cases with poor business consistency, while large language model based approaches face two major challenges: a high error rate in generated tests and insufficient code coverage. To address these issues, this paper proposes TestLoter, a logic-driven test generation framework. The core contributions of TestLoter are twofold. First, by integrating the structured analysis capabilities of white-box testing with the functional validation characteristics of black-box testing, we design a logic-driven test generation chain-of-thought that enables deep semantic analysis of code. Second, we establish a hierarchical repair mechanism to systematically correct errors in generated test cases, significantly enhancing the correctness of the test code. Experimental results on nine open-source projects covering various domains, such as data processing and utility libraries, demonstrate that TestLoter achieves 83.6% line coverage and 78% branch coverage. Our approach outperforms both LLM-based methods and traditional search-based software testing techniques in terms of coverage, while also reducing the number of errors in the generated unit test code.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"84 ","pages":"Article 101348"},"PeriodicalIF":1.8,"publicationDate":"2025-07-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144721370","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A methodology for empirical complexity analysis based on Newton’s polynomial interpolation","authors":"Rafael Fontes Sumitani,&nbsp;Lucas Victor da Silva Costa,&nbsp;Frederico F. Campos,&nbsp;Fernando Magno Quintão Pereira","doi":"10.1016/j.cola.2025.101347","DOIUrl":"10.1016/j.cola.2025.101347","url":null,"abstract":"<div><div>A cost model is a function that relates how often each part of a program runs depending on its inputs. Cost models can be derived automatically via the observation of counters: instrumentation that tracks execution of program operations. This paper defines Newton Counters: counters that can be described via a polynomial ranging on a single program input variable whose value can be read in constant time. Additionally, it shows that Newton Counters are prevalent in actual codes. Motivated by this observation, the paper introduces a methodology to derive automatic cost models. Said methodology combines static code analyses with interpolation via Newton’s divided difference method. This approach is currently available as a tool, <span>Merlin</span>. The effectiveness of this tool is demonstrated on 949 executable C programs taken from the <span>Jotai</span> collection, and on <span>genann</span>, a neural network library.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"84 ","pages":"Article 101347"},"PeriodicalIF":1.7,"publicationDate":"2025-07-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144703220","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Mind the gap: The missing features of the tools to support user studies in software engineering","authors":"Lázaro Costa ,&nbsp;Susana Barbosa ,&nbsp;Jácome Cunha","doi":"10.1016/j.cola.2025.101345","DOIUrl":"10.1016/j.cola.2025.101345","url":null,"abstract":"<div><div>User studies are paramount for advancing research in software engineering, particularly when evaluating tools and techniques involving programmers. However, researchers face several barriers when performing them despite the existence of supporting tools. We base our study on a set of tools and researcher-reported barriers identified in prior work on user studies in software engineering. In this work, we study how existing tools and their features cope with previously identified barriers. Moreover, we propose new features for the barriers that lack support. We validated our proposal with 102 researchers, achieving statistically significant positive support for all but one feature. We study the current gap between tools and barriers, using features as the bridge. We show there is a significant lack of support for several barriers, as some have no single tool to support them.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"84 ","pages":"Article 101345"},"PeriodicalIF":1.7,"publicationDate":"2025-07-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144679654","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Investigating the energy consumption of C++ and Java solutions mined from a programming contest site","authors":"Sérgio Queiroz de Medeiros,&nbsp;Marcelo Borges Nogueira,&nbsp;Gustavo Quezado","doi":"10.1016/j.cola.2025.101341","DOIUrl":"10.1016/j.cola.2025.101341","url":null,"abstract":"<div><div>The current concern about global warming has led to an increasing interest in the energy efficiency of computer applications. Assuming power is constant, the general trend is that faster programs consume less energy, thus optimizing a program for speed would also improve its energy efficiency.</div><div>We investigate this tendency in a set of C++ and Java solutions mined from Code Submission Evaluation System (CSES), a popular programming competition site, where each solution must give the correct answer under a given time limit. In such context, we can consider that all correct solutions for a problem were written with a speed concern, but not with energy efficiency in mind.</div><div>We selected 15 problems from CSES and for each of them we mined at least 30 C++ and Java solutions, evaluating time and energy efficiency of each solution in at least two different machines. In our scenario, where there is a great diversity of programming styles, execution speed, and memory usage, we could confirm the general trend: faster programs consume less energy. Moreover, we were able to use ordinary least squares to fit a linear function, with good precision, that relates energy consumption of a program to its execution time, as well as to automatically identify programs with abnormal energy consumption. A manual analysis of these programs revealed that often they perform a different amount of allocation and deallocation operations when compared to programs with similar execution times.</div><div>We also calculated the energy consumption profile of sets of random C++ solutions for these 15 CSES problems, and we tried to associate each set with its corresponding CSES problem by using the energy consumption profiles previously computed for each one of them. By using this approach, we could restrict, for each set of random C++ solutions, the classification task to a subset of 7 CSES problems, a reduction of more than 50% in the search space.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"84 ","pages":"Article 101341"},"PeriodicalIF":1.7,"publicationDate":"2025-06-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144308119","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Python’s evolution on Stack Overflow: An empirical analysis of topic trends","authors":"Fengqi Hu,&nbsp;Weihao Xue,&nbsp;Siyuan Zhou,&nbsp;Ye Wang,&nbsp;Bo Jiang,&nbsp;Qiao Huang,&nbsp;Hua Zhang","doi":"10.1016/j.cola.2025.101340","DOIUrl":"10.1016/j.cola.2025.101340","url":null,"abstract":"<div><div>With the rapid development of information technology and changing programming practices, the demand for programming discussions on online Q&amp;A platforms is growing. This study analyzes over two million Python-related posts on Stack Overflow to identify core topics and challenges over fifteen years. By using a Gradient Boosting Decision Tree (GBDT) model to quantify post popularity, we objectively show what the hottest as well as the most disturbing topics related to Python are to users at different times. We find that: The domains most closely associated with Python are data processing and machine learning, while development environments as well as automation and testing are gradually increasing in popularity. Machine learning is the area that bothers users the most. Moreover, we found that some questions that confuse users can increase the popularity of related topics. These findings can help developers grasp the direction of the Python language so that they can better plan their personal learning and project development. Enterprises and organizations can also optimize resource allocation based on trends in hot topics for training, tool development, and technical support.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"84 ","pages":"Article 101340"},"PeriodicalIF":1.7,"publicationDate":"2025-06-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"144297313","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A novel framework for evaluating developers’ code comprehension proficiency through technical and non-technical skills","authors":"Divjot Singh,&nbsp;Ashutosh Mishra,&nbsp;Ashutosh Aggarwal","doi":"10.1016/j.cola.2025.101327","DOIUrl":"10.1016/j.cola.2025.101327","url":null,"abstract":"<div><h3>Context:</h3><div>Code comprehension is an essential software maintenance skill, where technical skills are often considered the primary benchmark for evaluating developers’ proficiency, overlooking the significant role of non-technical skills.</div></div><div><h3>Objective:</h3><div>Our work aims to propose a generalized framework for measuring developers’ code comprehension proficiency by integrating technical and non-technical skills, inspired by cognitive attraction networks, and conducting an empirical study to evaluate code comprehension proficiency based on selective skills.</div></div><div><h3>Methods:</h3><div>The generalized framework evaluates developers’ technical and non-technical skills separately using collected data and computes their respective indices to derive an overall measure of code comprehension ability, represented as the comprehension measure index (CMI). Additionally, an empirical study with 158 participants assessed technical skills, including code understanding, debugging, and completion, alongside non-technical skills such as problem-solving, emotions, long-term memory, belief, desire, intention, and commitment to compute their overall code comprehension proficiency.</div></div><div><h3>Results:</h3><div>Based on the obtained indices values related to technical and non-technical parameters, the study identifies multiple factors affecting participants’ performance, including lack of technical knowledge, reliance on guesswork, stress intolerance, lack of commitment and desire, difficulty understanding logic, inability to recall concepts, and check other contributing factors. To enhance our results K-means clustering is done to group the participants into three clusters according to their performance.</div></div><div><h3>Conclusion:</h3><div>Integrating technical and non-technical skills enables a more accurate assessment by addressing factors beyond technical expertise. The framework can help managers and tutors identify strengths and weaknesses, allowing task assignments that align with strengths of developers while addressing areas for improvement.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"83 ","pages":"Article 101327"},"PeriodicalIF":1.7,"publicationDate":"2025-04-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143895592","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The evolution of Lua, continued","authors":"Roberto Ierusalimschy ,&nbsp;Luiz Henrique de Figueiredo ,&nbsp;Waldemar Celes","doi":"10.1016/j.cola.2025.101326","DOIUrl":"10.1016/j.cola.2025.101326","url":null,"abstract":"<div><div>Lua is a scripting language created in 1993 in Brazil. We have reported in detail on the birth of Lua and its evolution until 2007. Here, we chronicle the evolution of Lua since then. In particular, we discuss in detail the evolution of global variables, the introduction of integers, and the implementation of garbage collection and finalizers, including deterministic finalization. We also comment on some landmark social developments in the history of Lua.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"83 ","pages":"Article 101326"},"PeriodicalIF":1.7,"publicationDate":"2025-04-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143834487","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Debugging in the Domain-Specific Modeling Languages for multi-agent systems","authors":"Baris Tekin Tezel ,&nbsp;Geylani Kardas","doi":"10.1016/j.cola.2025.101325","DOIUrl":"10.1016/j.cola.2025.101325","url":null,"abstract":"<div><div>In many cases, developers face challenges while implementing Multi-Agent Systems (MAS) due to the complexity of expanding software systems, despite the presence of numerous agent programming environments and platforms. To tackle this complexity, Model-driven Engineering (MDE) can be employed at a higher level of abstraction and component modeling before diving into MAS development, which helps alleviate the intricacies. Probably, the most effective method of incorporating MDE into Multi-Agent Systems (MAS) is to adapt Domain-Specific Modeling Languages (DSMLs) along with integrated development environments (IDEs). These tools make it easier to model the system and generate the necessary code for the development process. Although existing MAS DSML IDEs offer some control over systems modeled based on the language’s syntax and semantics, they lack built-in debugging support. This deficiency leads to uncertainty among agent developers about the accuracy of models prepared during the design phase. To address this issue, this study proposes a comprehensive debugging framework (MASDebugFW) that facilitates the design of agent components within modeling environments. The framework’s utilization commences with modeling MASs using a design language, and then converting these design model instances into a runtime model. Following that, the runtime model undergoes simulation using an integrated simulator specifically designed for debugging purposes. Additionally, the framework includes a simulation environment model and a control mechanism to manage the simulation process effectively. These features further enhance the debugging capabilities and overall functionality of MASDebugFW. Furthermore, we have qualitatively and quantitatively evaluated MASDebugFW, subjecting all obtained results to statistical analysis. The evaluation results show that, on average, the implemented framework reduces debugging time by around 45%, leading to more efficient debugging processes. Moreover, it significantly enhances bug detection and repair capabilities, as it increases the number of bugs fixed in the models by approximately 50%.</div></div>","PeriodicalId":48552,"journal":{"name":"Journal of Computer Languages","volume":"83 ","pages":"Article 101325"},"PeriodicalIF":1.7,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143480205","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}